Your money or your life? Public support for health initiatives during the COVID-19 pandemic.

While many Australians agree with the need for COVID-19 restrictions and lockdowns, a vocal minority of citizens loudly voice opposition to government restrictions, characterised as impositions on 'free speech' and impacting the Australian economy. National data from the 2020/21 Australian Survey of Social and Attitudes and Tasmanian survey data enable an examination of those who support or oppose health initiatives aimed at minimising the impact of the pandemic. The majority of Australians believed government imposed restrictions were 'about right', although women were more likely than men to be concerned about the health implications of COVID-19. Tasmanian data show that controlling for social background, those who prioritise the economy are less likely than those concerned about the health implications of COVID-19 to wear masks in public, less likely to be tested in case they have to self-isolate, less likely to check in, and less likely to be vaccinated. Alternatively, higher trust in university research, and trust in scientists as a source of information about COVID-19 was positively associated with attitudes and behaviour designed to reduce the transmission of COVID-19.

INTRODUCTION

In this research, I analyse Australian survey data collected at the national level and in the state of Tasmania, to examine empirically, some of the factors underlying non‐compliance with COVID safe behaviour. Like the rest of the world, Australians have been exposed to various strains of the COVID‐19 virus, with the first outbreak occurring in the state of Victoria on 25 January, 2020. The first wave occurred between March and April 2020, and Australia closed its borders to non‐citizens and non‐residents on 20 March, 2020 (Murphy and Karp 2020). While the Australian federal government is responsible for controlling borders and vaccine supply into the country, state and territory governments are primarily responsible for administering vaccines, regulating and policing rules relating to citizens within state and territorial boundaries, and instituting preventative measures aimed at preventing the spread of COVID‐19.

Australian states and territories have imposed a variety of restrictions upon their citizens aimed at restricting viral transmission. These include mandating the wearing masks in places of work, educational institutions, retail venues and public places, “social distancing” rules, capping the numbers of people who can congregate in public and private spaces, and home and hotel quarantine procedures. Governments have also limited the movement of those referred to as “non‐essential workers” when viral transmission rates were deemed high.

States and politics

Some Australian states experienced higher COVID‐19 infection rates and subsequently imposed more severe “lockdown” rules. Australian Federal Treasurer Josh Frydenberg claimed that his home state of Victoria experienced the longest periods of “lockdown” in the world during 2020 and 2021 (Frydenberg 2021), although Campbell (2021) disputes this claim, arguing Argentina's lockdown was longer. Either way, few would argue that Victorians have experienced by far the most severe and sustained restrictions in Australia during the COVID‐19 pandemic. New South Wales also experienced prolonged periods of lockdown in 2021 following the emergence of the Delta strain of the virus, along with the Australian Capital Territory, while the remaining states and territory governments imposed shorter period of lockdown and fewer restrictions on the movement of citizens. State‐based differences in acceptance of government restrictions may therefore be expected to vary in the survey responses, with resistance expected to be higher in Victoria and (to a lesser extent) NSW.

The political parties in government federally and in each state and territory may also have an influence on public attitudes. Australian state and territory governments of different political persuasions have tended to vary in their approach to restrictions on movement of their citizenry. In the two most populous states, the Liberal/National government in NSW was slower than other states to impose restrictions on the movement of citizens, preferring “track‐and‐trace” approaches to “lockdowns”, at least prior to the rapidly spiralling case numbers due to the Delta strain in 2021 (www.covid19data.com.au). The Victorian Labor government, on the contrary, was quicker to confine citizens to their homes, impose mask wearing and restrict access to non‐essential services.

The remaining states generally adopted more cautious approaches than NSW, regardless of the parties in power. Queensland (Labor) closed its borders quickly to other states experiencing COVID‐19, as did the Liberal‐led state of South Australia. The island state of Tasmania (Liberal) adopted a fortress approach, closing its borders early to the rest of the country (Council of the Ageing Tasmania, 2021), while a the Labor state of Western Australia remains closed until vaccination rates reach 80 per cent of people aged 16 and older (Shine, 2021).

Nevertheless, supporters of conservative parties are expected to be generally more concerned about the economic impacts of government restrictions upon business activities and employment, although the question of whether the influence of party affiliations varies across states when measuring COVID‐related attitudes and behaviour remains open. This research will examine the association between political party identification and attitudes and behaviour relating to COVID‐19.

Trust in science and government

The extent that one trusts governments, the advice of public experts and, relatedly, trust in science are important factors in vaccine uptake and compliance with public health advice. According to survey evidence from 113 countries collected by the Wellcome Global Monitor (2020), trust in science has increased during the pandemic, as “Globally, those who said they trust scientists ‘a lot’ rose from 34 per cent in 2018 to 43 per cent by the end of 2020.”

Bargain and Aminjonov (2020, 13) found “higher political trust is associated with a larger reduction in non‐essential mobility following the implementation of containment policies.” In fact, in cross‐national survey research that included Australia, Algan et al. (2021) found “trust in scientists is the critical determinant of societies' resilience in their fight against the pandemic,” while the influence of trust in government “is less clear‐cut”. However, they also found that when trust in science declines for individuals, compliance with public health advice also declines (Algan et al., 2021).

Research in the United States on the influence of ideology and trust on willingness to have non‐COVID vaccinations (i.e., flu, pertussis and measles) found conservatives “less likely to express pro‐vaccination beliefs than other individuals” along with those with “lower levels of trust in government medical experts” (Baumgaertner, Carlisle and Justwan, 2018, 1; 9). Baumgaertner et al. (2018) also maintain one's ideology influences who one trusts. In a recent study of public opinion relating to COVID‐19 in southern Europe, Lesschaeve, Glaurdic and Mochtak (2021) show that citizens generally prioritise saving lives over economic costs. However, they also found that “free‐market views also make people more accepting of higher casualties, as do fears that the instituted measures will lead to a permanent expansion of government control over society” (Lesschaeve, Glaurdic and Mochtak, 2021). Similarly, Biddlestone, Green and Douglas (2020) found those with individualist, as opposed to collectivist worldviews, were less likely to behave in ways that prevent the spread of COVID‐19.

Researching public attitudes on COVID‐19 in the United States, Hamilton and Safford (2021, 1) found that public trust in scientists varied according to party allegiances, with trust in science agencies substantially lower among those who identify with conservative political parties. Studying public attitudes on the COVID‐19 pandemic in Brazil, Coelho et al. (2021) examined how values and political orientations are linked to prioritising economic or health concerns. They found “cultural values such as embeddedness (e.g., tradition, conformity) and hierarchy (e.g., power) are linked to a higher mobility…cultures that highly endorse those values tend to follow lockdown measures less” (Coelho et al., 2021). Further, they found that the political left tended to be motivated by humanitarian concerns, in contrast to the right who prioritised the economy (Coelho et al., 2021). I expect that conservative party identifiers (i.e., Liberals and Nationals) will be more likely than those affiliated with progressive parties (i.e., Labor and Greens) to reject the advice of health authorities relating to COVID‐19, and that political party allegiances will cross‐cut trust in science relating to COVID‐19 public health initiatives in Australia.

Postmaterialist values and health

Inglehart (1997, 33–34) argues that values are formed in early adolescence, and remain relatively stable over the life course, but the economic and social conditions experienced during one's formative years influence the development of materialist or postmaterialist value priorities. Based on Maslow's (1954) hierarchy of needs, the postmaterialist value thesis suggests that basic physiological and physical needs must be satisfied before individuals shift their focus to higher‐order intellectual, social and aesthetic needs (Inglehart, 1997, 33). When physical safety and physiological needs are met, people tend to shift their priorities to needs relating to esteem, self‐expression and self‐realisation. The postmaterialist value thesis attributes generationally based value change to different levels of economic prosperity and physical security, so that in advanced industrial societies, postmaterialists are claimed to prioritise quality of life and self‐expression over economic growth (Inglehart, 1997, 33–36). As Inglehart and Baker (2000, 26) put it “these values tap an intergenerational shift from an emphasis on economic and physical security toward an increased emphasis on self‐expression, subjective well‐being, and quality‐of‐life concerns.”

Inglehart (1977, 1997) measures materialist and postmaterialist values through questions administered in national surveys, with value priorities typically operationalised using a 4‐item value index (described below). The postmaterialist value index is also a key component of Inglehart's survival/self‐expression value dimension, where survival values tend to be associated with poorer health outcomes (Inglehart and Baker 2000, 27). In more recent Swedish research, Petersen and Lindstrom (2010, 798) found materialists  less likely than postmaterialists to believe their own behaviour determines their health status, which they suggested “may signify a causal relationship between materialist values and sense of less health control and control over health behaviours.” This finding is consistent with a German healthcare study by Hajek and Konig (2020, 1), who argued that “the increased likelihood of preventive medical check‐ups in post‐materialistic women will be beneficial in decreasing the need for doctor visits for reasons of chronic illnesses.”

In a study of 42 European countries, Mackenbach (2014, 130) found “countries with higher scores on ‘self‐expression’ have better health behaviours or health outcomes,” while Roudijk et al's. (2017) research on self‐reported health in 51 countries shows that “within countries, there is a positive association between the survival/self‐expression dimension and SRH” [self‐reported health]. I expect postmaterialists to be more concerned than materialists about the health implications of COVID‐19, and less concerned than materialists about the economic implications of the pandemic and associated “lockdown” rules.

Study aims

Using data from Australian Survey of Social Attitudes, I consider the factors associated with attitudes toward the COVID‐19 pandemic at the national level. Drawing upon survey data from Tasmania, I examine related public attitudes in a state that has remained largely isolated from the rest of the country for much of the pandemic. The AuSSA data (collected from February to June 2021) enable consideration of state‐based differences in public priorities (i.e., greater concerns over the economic versus health implication of the pandemic; perceptions of the how restrictive “lockdowns” were). The Tasmanian survey, collected from September to October 2021, contains both attitudinal and behavioural questions.

The focus here is upon the following research questions: Are Australians more concerned about the economic or health impacts of COVID‐19? How does prioritising the economy or health influence COVID‐19‐related attitudes and behaviour? How are social and political background and state of residence associated with public attitudes toward the impact of COVID‐19? Are “materialists” more concerned than “postmaterialists” about the economic impacts of COVID‐19? To what extent are trust in science and trust in government associated with attitudes toward COVID‐19 restrictions?

Hypotheses

Men are less likely than women to comply with COVID‐related restrictions imposed by governments.

Older Australians are more concerned than younger Australians about the health impact of COVID‐19.

Residents of Victoria and New South Wales are more concerned about “lockdowns” than those living elsewhere.

Liberal and National party identifiers are more likely than Labor identifiers to be concerned about the economic impacts of COVID‐19.

Trust in science and in government are positively associated with following the advice of health authorities.

Postmaterialists are more concerned about the health than the economic impacts of COVID‐19.

DATA AND METHOD

National data are from the “2020 Australian Survey of Social Attitudes” (AuSSA), a mail survey administered to Australians aged 18 and older by the Australian Consortium for Social and Political Research Incorporated (McNeil et al., 2021). AuSSA data collection was originally planned to commence in 2020 but delayed until 2021. The Australian Electoral Commission randomly selected 5000 citizens from the electoral roll. Potential participants were contacted with a pre‐notification letter in late February 2021, followed a week later with the survey package, then a reminder postcard 1 week later. After 6 weeks, a replacement survey package was sent to those who had not responded, followed by a final reminder card 1 week later. Data collection occurred between 22 February and 9 June 2021. In total, 1162 completed questionnaire responses were received, with 272 not eligible, for a response rate of 25 per cent [i.e., 1162/(5000–272)].

The Tasmania Project survey is conducted by the Institute for Social Change at the University of Tasmania. These cross‐sectional data are from a nonrandom sample of Tasmanians aged 18 years and older, collected between 24 September and 3 October 2021 (n = 1200). Tasmanian residents were recruited to the project by social media advertising, through University of Tasmania websites, and in newspapers, radio and television. Data were collected using an online self‐administered survey, although paper and telephone surveys were also available for those without Internet access. Face‐to‐face recruitment strategies were employed by the project team in rural and regional areas in addition to online recruitment.

This study has certain limitations. While the Australia Survey of Social Attitudes is a probability sample of Australian adults (N = 1162) and therefore useful for making estimates at the national level, a larger national sample would have enabled more fine‐grained analysis at the state and territory level. As a mainly online survey, Tasmanian data were not collected using probability sampling methods. While the Tasmanian data were weighted to match Tasmanian Census data for age, sex and education, the Tasmanian estimates may not be representative of the population of Tasmanian adults.

Dependent variables

Two dependent variables are examined in the national analyses. The first is derived from the question: “Many people have been experiencing the effects of COVID‐19 pandemic (Coronavirus) in a variety of ways. Are you personally more concerned about the health impacts or the economic impacts of the COVID‐19 pandemic?” (five response categories: much more concerned about the health impacts to much more concerned about the economic impacts). The second variable is from the question “In your opinion, were the ‘lock down’ rules imposed by state and federal governments too restrictive or not restrictive enough?” (five response categories: definitely too restrictive to definitely not restrictive enough). Given the ordinal structure of these variables, they are analysed using ordinal logistic regression analysis (with STATA, version 16).

Variables analysed using the Tasmanian data measure attitudes and behaviour relating to the pandemic. Once again, several of these variables have an ordinal structure and are analysed using ordinal logistic regression analyses. The first four of these variables have 5‐point Likert (i.e. strongly agree to strongly disagree) response categories. The first three variables “If I showed mild symptoms of COVID‐19, I would be tested,” “I would avoid being tested for COVID‐19 so I do not have to self‐isolate while waiting for results” and “I would wear a face mask in public to limit the spread of COVID‐19” are (for many Tasmanians given the low virus cases and short periods of lockdown) attitudinal questions, as they relate to what were for most Tasmanians, hypothetical questions when these data were collected. However, the remaining Likert question can be classified as behavioural, as it relates to an actual behaviour “I do NOT always check in to public venues for contact tracing.” Two other variables also measure behaviours: “have you been partially or fully vaccinated against COVID‐19?” (yes/no) and “Are you… (responses fully vaccinated [two doses]; partly vaccinated [one dose]).” These dependent variables are analysed using binary logistic regression analysis.

Independent variables

Several dummy independent variables are examined with the AuSSA data. Australian research by Enticott (2021) found that younger males living in major cities were less likely to engage in physical distancing, self‐quarantining if unwell, or not testing for COVID‐19 when symptomatic. Edwards et al. (2021, 8) found that Australian “females, those with lower levels of household income and living in disadvantaged areas were associated with increased likelihood of vaccine resistance or hesitancy.” However, they also found that “in contrast to previous research, younger people were not less likely to intend to get vaccinated than those aged 35–44 years.”

I operationalise respondent sex, and after some preliminary analyses, an older age group (i.e., respondent sex, men = 1; aged 70+ = 1), as well as tertiary degree = 1; currently in paid work = 1; living in a “big city” = 1 and low personal income <$30,000 = 1. Australian states are scored 1 with Victoria as the reference group, while territories are excluded due to very low responses. Political party identification has Labor as the reference category; while self‐rated fair or poor health is scored 1, with excellent, very good, or good health as the reference category. A variable was also constructed from a question in the AuSSA measuring loneliness during COVID “Since the COVID‐19 pandemic restrictions began I have been, or was, more lonely than usual” (response strongly agree + agree scored 1; other responses scored 0).

Postmaterialist values are operationalised following Inglehart (1977, 28), where responses to two questions “which do you think should be Australia's highest priority, the most important thing it should do” and “which one do you think should be Australia's next highest priority, the second most important thing it should do?” are combined. That is, any combination of the responses “maintain order” and “fight rising prices” are classified as materialist; any combination of responses “give people more say in important government decisions” and “protect freedom of speech” are classified as postmaterialist. Any other combination of responses are classified as having mixed values.

Several items measure various forms of trust. The first measuring interpersonal trust (Tranter & Skrbis, 2009) is “Generally speaking, would you say that most people can be trusted, or that you can't be too careful in dealing with people?” (response categories numbered 1 (you can't be too careful) to 5 (most people can be trusted). Categories 4 and 5 are scored 1 for high trust).

Four variables are derived from questions measuring institutional trust with scales ranging from 0 (no trust at all) to 10 (complete trust): “On a scale of 0–10, how much do you personally trust each of the following institutions? 0 means you do not trust an institution at all, and 10 means you trust it completely” (University research centres; the news media; business and industry; and the federal parliament). Finally, an AuSSA question measures trust in scientists as a sources of information “How much trust do you have in scientists as a source of information about…?” While a range of topics were listed: weather forecasting; forest management issues; vaccines; climate change; and genetically modified crops (five response categories: a great deal of trust to no trust at all), vaccines are of particular interest here, and are included in the models. Notably, the question does not ask respondents about COVID‐19 vaccines, but about vaccines in general.

For the Tasmania Project data, I examine the following independent variables (respondent sex: men = 1), age group dummy variables Aged 65+ (referent: Aged 18–29 = 1; Aged 30–49 = 1; and Aged 50–64 = 1), university degree or higher degree, Aboriginal or Torres Strait Islander status, and region (i.e., southern, northern, northwest and west). For the Tasmanian data, political party identification contrasts non‐party identifiers = 1 and other party identifiers = 1 with Labor, Liberal and Greens identifiers. This is because when analysing some of the key dependent variables such as vaccination status, most variation in responses occurred between the established parties compared with “other” party identifiers, or non‐partisans.

Institutional trust was measured with two key variables: “I trust government and officials to provide reliable information about COVID‐19” and “I trust scientists to provide reliable information about COVID‐19” (5‐point Likert responses: strongly agree + agree = 1; other responses = 0).

ANALYSES

The responses shown in Figure 1 indicate that a greater proportion of Australians (22 per cent) are much more concerned about the health impacts of the COVID‐19 virus than the economic impacts (8 per cent). Nevertheless, by far the largest proportion, 50 per cent of respondents are “about equally concerned” about health and the economy. On the other hand, 8 per cent maintain “lockdown” rules were “definitely too restrictive,” compared with only 4 per cent who believed such rules were “definitely not restrictive enough.” However, once again, the modal response indicates that for 65 per cent of all respondents, lockdown rules were ‘about right’.

FIGURE 1

Attitudes toward COVID‐19 concerns over health versus economy and lockdown rules

A crosstabulation of the association between the two variables shown in Figure 1 indicates a clear pattern of association between perceptions of the severity of lockdown rules and prioritising the economy over health (see Figure 2). Those who believe the lockdown rules were too restrictive (41 per cent) are far more likely to be concerned about the economic impact of COVID‐19 than those for whom the lockdown was not restrictive enough (5 per cent). Alternatively, 50 per cent of respondents who claimed the lockdown was not restrictive enough were more concerned about the health impacts of COVID, compared with only 19 per cent of those who claimed the rules were too restrictive.

FIGURE 2

Health vs economic impact of COVID‐19 by perceptions of “lockdown” rules

In Table 1, I consider how a variety of social background, political party identification, value orientations, Indigenous status and self‐rated health measures are associated with prioritising the economy or health and attitudes toward the lockdown. Table 1 shows men (23 per cent) to be more concerned than women (11 per cent) about the economic impacts of the pandemic, while women (38 per cent) are more concerned than men (30 per cent) about the health impacts. Age‐based differences are extant, but there is no clear decline or increase in attitudes with age. Rather, the youngest and older groups are most concerned about health, while those aged 30–49 are most concerned about the economy.

TABLE 1

“ How restrictive were ‘lockdown’ rules and concerns over the health or economic impacts of COVID‐19” (per cent)

	Health	About equal	Economy	Too restrictive	About right	Not restrictive
Men	30a	48a	23a	24a	60a	15a
Women	38b	51a	11b	15b	70b	15a
18–29	36ab	51ab	14ab	18a	63a	19a
30–49	31b	45b	24c	20a	69a	11a
50–64	28b	54a	18bc	21a	63a	16a
65+	41a	48ab	12a	18a	66a	16a
Non‐tertiary	32a	52a	16a	20a	63a	17a
Degree	36a	46b	18a	19a	69a	12b
Labor	43a	45a	12a	9a	77a	14a
Coalition	24b	52a	24b	28b	59b	13a
Greens	50a	42a	9a	13 ac	70ab	17a
No party	32c	52a	16a	19c	64b	17a
Materialists	30a	53a	17a	18a	66a	16a
Mixed	34a	49a	17a	18a	67a	15a
Postmaterialists	35a	49a	17a	29b	56b	15a
Personal income $0–30 K	41a	48a	11a	19a	64a	17a
Personal income $30 K+	31b	51a	23b	20a	66a	15a
Aboriginal/Torres Strait Is	36a	47a	17a	14a	50a	36a
Other	34a	50a	16a	20a	66b	14b
Health – fair/poor	39a	49a	12a	21a	55a	25a
Excellent/very good/good	32a	50a	18a	19a	68b	13b
All	34	50	17	20	65	15

Note: Compared down each column, different letters signify statistical significance at p < .05.

Tertiary education has a very weak association with the health versus economy question; however, there are notable differences based on political identification. Coalition identifiers (24 per cent) are less likely than Labor identifiers (43 per cent) to be concerned about the health implications of COVID‐19, but are far more likely than Labor identifiers to be concerned about the economic impact of the virus. Income inequality is also associated with attitudes toward health and the economy, with people earning $30,000 or less (41 per cent) more likely than those on higher incomes (31 per cent) to be concerned about health, and less concerned about the economy. Somewhat surprisingly, there is very little variation in these data on the basis of value orientations.

Finally, I consider how self‐rated health is associated with attitudes toward the pandemic. While self‐rated health is a rather blunt tool, it has been shown to be a useful proxy for measuring actual health (Burstrom and Fredlund 2001; Doiron et al., 2015; Tranter & Donoghue, 2016). The assumption here is that those who experience poor health will be more concerned about health impacts, because they may be more vulnerable to becoming seriously ill during the pandemic. There is some evidence to support this hypothesis, with 39 per cent of those who rate their health as “fair” or “poor” concerned about the health impacts of the pandemic, compared to 32 per cent of people with “excellent”, “very good” or “good” health (p < .01).

Turning to the question were the “lockdown” rules imposed by state and federal governments too restrictive or not restrictive enough? men (24 per cent) were more likely than women (15 per cent) to claim lockdown restrictions were too restrictive, although there were no significant differences according to age group. Tertiary graduates were less likely (12 per cent) than non‐graduates (17 per cent) to view lockdown rules as not restrictive enough. Once again, political party identification was associated with substantial attitudinal differences in the lockdown rule question. Coalition identifiers (28 per cent) were far more likely than Labor (9 per cent), Greens (13 per cent) or those who do not identify with any political party (19 per cent) to claim lockdown rules were too restrictive, while Labor identifiers were least likely to claim these rules were “about right”.

Unexpectedly, postmaterialists (29 per cent) were more likely than materialists (18 per cent) or those with “mixed” values (18 per cent) to view lockdown rules as restrictive, and postmaterialists were least likely to claim the rules were “about right”. Once again, self‐assessed health status is associated with attitudes on the lockdown. Aboriginal and Torres Strait Islander people (50 per cent) were less likely than others (66 per cent) to agree that lockdown restrictions were “about right”, and more likely (36 per cent) than other Australians (14 per cent) to claim lockdowns were not restrictive enough. These findings may reflect the poor health status of many Indigenous Australians and possibly also indicate lower vaccination uptake among Indigenous peoples.

Regression analyses

While Table 1 indicates attitudinal variation on the bases of several variables (e.g., respondent sex, age, party identification, ATSI status, income and health), these independent variables are themselves correlated in various ways. In order to control for correlations between independent variables, I employ regression models in the following regression analyses. While variables with three categories are presented for the bivariate analyses in Table 1, the full range of variation in the health versus economy and lockdown variables is modelled in Table 3, using ordinal logistic regression analysis (i.e., as these dependent variables have an ordinal structure). Odds ratios are also reported.

TABLE 3

Crosstabulations of COVID‐19 pandemic attitudes and behaviour with social and political background (per cent)

	No test	No check in	No isolation	No mask	No‐vax	One dose only
Men	10a	23a	8a	19a	17a	9a
Women	13a	10b	7a	9b	10b	24b
18–29	18a	25a	2a	21a	37a	51a
30–49	18a	18a	11b	15ab	24b	21b
50–64	9b	11b	8bc	9c	9c	12c
65+	5c	10b	5 ac	12bc	3d	7d
Degree	11a	14a	8a	9a	8a	14a
Non‐tertiary	12a	16a	7a	14b	18b	19a
Aboriginal or TSI	17a	4a	15a	17a	45a	50a
Other	11a	16b	7b	13a	15b	17b
Liberal	11ab	13ab	10a	11a	1a	18ab
Labor	8ab	8b	3b	8a	11b	24b
Greens	7b	15a	10a	1b	12b	23b
No party	13a	18 ac	7ab	20c	20c	13a
“Other” party	23c	25c	3ab	27c	58d	5a
Region – South	8a	14a	4a	14a	15a	16a
North	7a	16a	17b	11a	9b	10b
Northwest/West	20b	17a	7ab	14a	21c	25c
Health limits my lifestyle	7a	16a	1a	13a	26a	25a
Health less limiting	12a	15a	7a	13a	15b	16a
All	11	15	7	13	16	17

Note: Compared down each column, different letters signify statistical significance at p < .05

Table 2 regresses a range of relevant independent variables on the two dependent variables. The ordinal logistic regression analyses suggest that even after controlling for social background, political partisanship, value orientations and measures of trust in university research and in business, men (OR 1.4) are significantly more likely than women to be concerned about the economy over health. People older than 70 (OR 0.5) show lower odds than younger Australians regarding concern about the economic impact of COVID‐19 (Figure 2). Those living in large cities (OR 1.6) are more concerned about the economy than health, although there are no significant state‐based differences for this dependent variable. However, Coalition identifiers (OR 1.6) are more likely than Labor partisans to be concerned about the economy as opposed to the health impact of COVID‐19. The odds for Greens and non‐partisans show no significant differences from the Coalition reference category at the 95 per cent level.

TABLE 2

Ordinal logistic regression analyses of COVID‐19 pandemic attitudes (odds ratios)

	Economy vs health	Lockdown restrictive
Men	1.4*	1.3^
Women	1	1
Aged <70	1	1
Aged 70+	0.5***	0.9
Non‐tertiary	1	1
Degree	0.99	1.2
Live in big city	1.6*	1.4^
Live outside city	1	1
Personal income $0‐30 K	0.8	1.1
Personal income $30 K+	1	1
New South Wales	0.96	0.5**
Victoria	1	1
Queensland	1.1	0.7
South Australia	0.8	0.4**
Western Australia	0.9	0.6*
Tasmania	0.99	0.7
Labor	1	1
Coalition	1.6*	1.5^
Greens	0.8	0.97
No party	1.3	1.2
Materialists	1	1
Mixed	1.1	1.03
Postmaterialists	1.3	1.9*
ATSI	1.05	0.4^
Other	1	1
Health–excellent/v. good/good	1	1
Fair/poor	0.7*	0.7^
High trust in others	0.9	1.5*
Less trust in others	1	1
Trust university research	0.86***	0.83***
Trust business	1.13**	1.07
Trust vaccine science	1	1
Do not trust vaccine science	1.1	1.6**
Economy most important national issue	2.4***	1.4^
Other issue more important	1	1
More lonely during COVID	0.9	1.7**
Not more lonely during COVID	1	1
Lockdown too restrictive	3.4***	‐
Lockdown not too restrictive	1	‐
N	(855)	(858)

Note: ^p < .1; *p < .05; **p < .01; and ***p < .001.

Those with poor or fair health (OR 0.7) are less concerned than those with better self‐rated health about the economic impacts of COVID 19, while the two scales measuring trust in university research and in business show highly significant associations with the economy versus health dependent variable. Those who score higher on the trust in university research scale (OR 0.86) are less likely to be concerned about the economic impact of COVID‐19, while trust in business (OR 1.13) shows the opposite association – the odds of concern over the economy increase with trust in business. Believing that “the economy” is the most important national issue increases the odds of concern over the economic implications of COVID‐19. Finally, those who perceive lockdown rules as too restrictive (OR 3.4) are more concerned about the economy than health in relation to COVID‐19.

Results for the second dependent variable show considerable state‐based differences, particularly between Victoria (reference category) and the other states (ACT and NT are removed due to very small sample size). Victorians are more likely than people living in all other states to view the lockdown rules are too restrictive, perhaps unsurprisingly given Victorians endured by far the longest lockdowns to date. Once again, Coalition identifiers (OR 1.5; p < .01) had higher odds than Labor identifiers in assessing lockdown rules as restrictive, while similar effects were apparent to the first dependent variable for trust in universities. Those who do not trust vaccine science (OR 1.6) and those who experienced loneliness to a greater extent during COVID‐19 (OR 1.7) were more likely to believe the lockdown has been too restrictive.

Tasmanian analyses

I examine six dependent variables that measure various attitudinal and behavioural responses to the COVID‐19 pandemic in Tables 3 and 4. These variables measure “If I showed mild symptoms of COVID‐19, I would be tested” (No test); “I do NOT always check in to public venues for contact tracing” (No check in); “I would avoid being tested for COVID‐19 so I do not have to self‐isolate while waiting for results” (No Isolation); “I would wear a face mask in public to limit the spread of COVID‐19” (No mask); “have you been partially or fully vaccinated against COVID‐19?” (No‐vax) and “Are you… fully vaccinated (2 doses); partly vaccinated (1 dose)” (1 dose only).

TABLE 4

Binary and ordinal logistic regression analyses of COVID‐19 pandemic attitudes and behaviour (odds)

	No test	No check in	No isolation	No mask	No‐vax	One dose only
Men	0.5	2.4*	1.4	1.8	2.0	0.3*
Women	1	1	1	1	1	1
Aged 65+ referent	1	1	1	1	1	1
Aged 18–29	4.6^	3.4	0.3	1.4	16.0***	18.8***
Aged 30–49	2.4^	2.0	2.3	1.2	10.6**	3.3*
Aged 50–64	1.3	1.2	1.9	0.7	3.8^	1.6
Non‐tertiary	1	1	1	1	1	1
Degree	1.2	0.9	1.4	0.6	0.5	0.6^
Aboriginal or TSI	0.6	0.1*	3.3	0.7	5.1*	2.1
Other	1	1	1	1	1	1
Major party or Greens ID	1	1	1	1	1	1
No political party identification	1.05	1.1	0.7	2.2	1.3	0.4
Other party	4.4*	1.9	0.2^	4.6*	2.0	0.3
Region (South referent)	1	1	1	1	1	1
North	0.6	1.5	5.1**	1.1	1.8	0.7
Northwest/West	2.9*	1.1	0.95	0.7	1.8	1.7
Trust government	1.7	1.6	1.3	0.7	0.5	1.04
Less trust in government	1	1	1	1	1	1
Trust scientists	0.6	0.2**	0.3*	0.4	0.2**	0.8
Less trust in scientists	1	1	1	1	1	1
Health limits my lifestyle	0.5	1.2	0.2	0.9	4.5^	2.9
Health less limiting	1	1	1	1	1	1
Agree economy greater concern than health for impact of COVID	4.5***	2.5*	8.2***	3.3*	2.8*	1.1
Unsure, disagree	1	1	1	1	1	1
N	(1037)	(1037)	(1037)	(1037)	(1033)	(959)

Note: ^p < .1; *p < .05; **p < .01; and ***p < .001.

The crosstabulation results in Table 3 show that men (23 per cent) are less likely than women (10 per cent) to check in to public venues, and less likely to wear a mask in public. Men are less likely than women to have been vaccinated, although fewer men (9 per cent) than women (24 per cent) have had only one vaccine dose. In other words, while men are less likely than women to be vaccinated, among the vaccinated, men are more likely than women to have been double‐vaccinated. There are some predictable age‐based results here. Younger Tasmanians are less likely than others to be tested if they experience mild COVID‐19 symptoms. Younger people are also less likely to have been vaccinated than older Tasmanians, and are less likely to always check in, to wear a mask, to be vaccinated and to have been double‐vaccinated. Tertiary graduates are significantly more likely than non‐graduates to wear a mask and to be vaccinated. Aboriginal and Torres Strait Islanders (ATSI) status is also associated with several of these dependent variables in the bivariate case. ATSI respondents appear are more likely than non‐Indigenous Tasmanians to check in to public venues, but are less likely to isolate if they experience COVID symptoms. Importantly, ATSI (45 per cent) are much more likely other Tasmanians (15 per cent) to be unvaccinated, and more likely (ATSI 50 per cent; others 17 per cent) to be single‐ rather than double‐vaccinated.

Political party identity effects are also apparent in the Tasmanian sample. Only 1 per cent of Liberal identifiers in the Tasmanian survey are unvaccinated, compared with 11 per cent of Labor, 12 per cent of Greens and 20 per cent of non‐identifiers. The very high uptake among the Liberal party identifiers may reflect the influence of the incumbent Liberal state government upon party followers. Regional differences are also apparent, with those in the Northwest of the state less likely to be tested and more likely to be unvaccinated, perhaps reflecting rural remoteness and problems accessing testing facilities and vaccination outlets.

The binary logistic regression analyses in Table 4 indicate that controlling for social background and other independent variables, men are less likely than women to check in to public venues, but more likely than women to have had two doses of the vaccine. Those who identify with “other” political parties are far less likely to be tested (OR 4.4), and less likely to wear a mask in public (OR 4.6) compared with those who identify with the Liberal, Labor or Greens parties. Regional differences are again important in the regression models, with those in the Northwest (OR 4.4) less likely than those living in Southern Tasmania to be tested if they have mild symptoms, while Northerners (OR 5.1) are less likely than Southerners to avoid testing, so they do not have to self‐isolate while waiting for results.

Institutional trust is important, particularly for vaccine uptake and checking in. While trust in government is only associated weakly with the dependent variables, trust in scientists has a strong association with three COVID safe indicators. Trusting scientists is positively associated with checking in, and with not avoiding isolating. Trusting scientists is also associated with increased odds of being vaccinated. Finally, among Tasmanians, even with a range of control variables in the regression models, those who are more concerned about the economic impact of COVID‐19 are less likely than those who prioritise health to engage in a range of public health initiatives designed to limit the spread of COVID‐19, with large odds for not isolating (OR 8.8) and not being tested (OR 4.5).

DISCUSSION

This research, based upon a nationally representative survey of Australian adults, and a survey of Tasmanian adults, shows that the majority of Australians are equally concerned about the health and economic implications of the COVID‐19 pandemic. Two‐thirds of the national sample held a Goldilocks view of the severity of the state government‐imposed lockdowns as neither too restrictive nor not restrictive enough, but “about right”. Not surprisingly, Victorians were the most likely to view the lockdowns as too restrictive, although they did not differ from other states on the economy versus health question. However, men tended to be more concerned than women about the economy over health, with men also more likely than women to perceive the lockdowns as too restrictive. Politically, identifying with the Liberal or National parties compared with Labor or the Greens was associated with less concerned over health, and greater concern over the economy during the pandemic. Those with conservative political leanings were also more likely to view lockdowns as “too restrictive”.

Notably, this research supports studies in other countries that highlight the importance of institutional trust for adhering to public health initiatives during the COVID‐19 pandemic (e.g., Safford, Whitmore and Hamilton 2021). Lorenzoni and Pidgeon (2006, 85) argue that trust in political institutions “reflects people's confidence in both the expertise and actions of agencies and institutions that initiate and control risk.” The United States has witnessed a decline in public trust in social institutions in recent years, including trust in government, partly due to the COVID‐19 pandemic (Edelman Trust Barometer 2021), and trust in a range of science disciplines and acceptance of scientific findings has become increasingly polarised over time (van der Linden and Lewandowsky 2015). Hamilton and Safford (2021, 1) found “the COVID‐19 pandemic has been marked by political divisions in U.S. public trust of scientists,” while trust in science agencies decreased substantially among those who identify with conservative political parties.

In Australia and New Zealand, trust in state and federal governments increased strongly during the COVID‐19 pandemic, as did trust in public health scientists (e.g., Goldfinch et al., 2021). Rejection of science, or suspicion of the veracity of scientific research is associated with a tendency to question or reject “expert advice”, with implications for a range of public concerns, from vaccine hesitancy to acceptance of anthropogenic climate change (Lewandowsky, Gignac and Oberauer 2013). Public trust in science is particularly important, as I found trust in university research and in scientists more broadly to be associated with a greater likelihood of prioritising health over the economy during the COVID‐19 pandemic, and of accepting public health initiatives, such as lockdowns. I also found that trust in science and scientists is far more important than trust in government when it comes to following public health initiatives designed to impede the spread of COVID‐19.

Similar to research on climate change (Lewandowsky et al., 2016), one's values and worldviews play an important role in understanding who follows the advice of health authorities regarding the pandemic. In Australia, protestors have repeatedly gathered in several cities across the country to rally for “freedom” from the lockdowns and other restrictions imposed by state governments (Dexter, 2021). As Lewandowsky et al. (2016) put it, when science comes up against “people's lifestyle or worldviews, or that impinge on corporate vested interests, the public response can be anything but favourable.” While Australia is a relatively postmaterialist country (Tranter and Wester, 2003; Western and Tranter, 2001), expectations that postmaterialists should be more concerned than materialists about the health implications of COVID‐19 were not borne out in the findings. The association between postmaterialist values and the health/economy dependent variable was weak and nonsignificant.

However, even more unexpectedly, postmaterialists were more concerned than materialists about the severity of the “lockdown” rules. These findings suggest that in Australia (and perhaps also in the United States), Inglehart's postmaterialist index is (at least partly) measuring concerns over the right to express “free speech” for some citizens, including anti‐lockdown demonstrators (Breakley, 2021). Further analysis (not shown here) indicates that those who prioritise the AuSSA survey item “protect freedom of speech” are significantly more likely than others to claim lockdown rules were too restrictive. These apparently right‐leaning Australians appear to be very different to the postmaterialists envisaged by Inglehart (1997; Inglehart and Baker, 2000), who tended to support progressive political issues such as the environment, women's rights, civil rights, sexuality and pro‐Green politics. These findings suggest that Inglehart's short value index is potentially flawed as a measure of postmaterialist values in contemporary Western democracies, where calls for “free speech” have become a rallying cry for many with conservative, and, in some cases, extremist political agendas.

Those who prioritise the economic implications of COVID‐19 are less likely than those concerned about health to be tested for the virus when they experience symptoms, and less likely to “check in” when visiting retail or other venues, to self‐isolate, to wear masks in public or to be vaccinated. These findings hold, even when controlling for social background variables, for political party allegiances and trust in scientists, highlighting the importance of accounting for worldviews in public health campaigns. As Goldfinch, Taplin and Gould (2021, 1) maintain “building trust is important for governments implementing difficult policy responses during a crisis.” Conspiracist thinking around the COVID‐19 pandemic builds resistance to following scientifically informed public health advice (Douglas 2021). In the United States, COVID‐19 conspiracy theories have been found to be associated with less concern over the risks posed by the pandemic, lower likelihood of adopting preventive measures to stop the spread of the virus, and greater likelihood of refusing vaccines (Romer and Jamieson 2020). Lewandowsky (2021) suggests “the best way to control conspiracy theories is to prevent them from spreading,” while also recommending “debunking after people have become familiar with a conspiracy theory.” Nevertheless, attempting to change the minds of people with strong conspiratorial tendencies may prove difficult.

While imposing restrictions upon citizens is necessary when attempting to combat a global pandemic, “stick” approaches can only take us so far. Vaccine uptake in Australia has been high, although pockets of resistance remain. Similar to climate change, it is important to attempt to understand the reasons behind resistance to following scientific advice on COVID‐19. It is necessary to focus on engaging with those who are hesitant, rather than focusing attention upon the relatively small number of people who loudly deny the scientific evidence of vaccination efficacy and the advice of health authorities – such as anti‐lockdown protestors. Further, in Australia, migrants who have experienced authoritarian regimes in their countries of origin, and Indigenous people with a long history of oppression in their own country understandably have low levels of trust in government authorities and are hesitant to engage with public health initiatives such as COVID‐19 vaccinations. Vaccination rates for Aboriginal and Torres Strait Islander people lag the national average (Briggs and Truu 2021). Indigenous Australians and migrant communities are likely to benefit from additional resources and government‐funded information campaigns delivered by their own community groups. Further research is needed to understand how best to engage with people who are authentically sceptical of vaccines and COVID‐safe practices, as opposed to those who deny outright, the scientific advice and the benefits of following public health initiatives.

AUTHOR CONTRIBUTION

BKT is the sole author and responsible for all writing and data analysis.

CONFLICT OF INTEREST

No perceived conflict of interest.